Answers are grounded in this paper and cite their sources.
The moments method is an attractive non-intrusive method, where estimating statistical moments and reconstructing the probability density function (PDF) are crucial to this approach. However, existing methods for estimating statistical moments face challenges in balancing efficiency and accuracy. Furthermore, the flexible distributions used to construct the PDF may have limited applicability for certain performance functions. In this paper, a full-domain moment method, which is based on a new point estimate method (PEM), monotonic transformation, and the normal inverse Gaussian (NIG) distribution, is proposed. First, an equivalent performance function is constructed through a properly designed monotonic transformation, which maintains the same failure probability as the original performance function but shifts the statistical moments into the applicable domain of the NIG distribution. Then, a new PEM, which combines the bivariate adaptive hybrid dimension reduction method (B-AH-DRM) and the Kriging surrogate model, is employed to estimate the first four central moments of the equivalent performance function. Based on the estimated first four central moments and the NIG distribution, the PDF of the equivalent performance function is constructed, and the failure probability is then calculated. Finally, several examples are used to validate the effectiveness of the proposed method in structural reliability analysis.
No takes yet. Share an insight, caveat, or question.
Fan et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: